GLAM Labs Projects

Coding da Vinci was – back in 2014 – the first German cultural data hackathon for open cultural data. It is a platform for innovative projects that release the creative potential of digital cultural heritage through creative projects created through a hackathon format.

Presentation video by Philippe Genêt from the German National Library (DNB) and further information at the website below. 

How Meemoo used AI to create and enrich descriptive metadata at the Flemish Institute for Archives.

Presentation video by Ellen Van Keer from Meemoo and further information available through the article below.

Presentation video from Stephanie Nitsche from the German National Library about how to make API requests and how to use the 'data dump' apps, see further information below from the DNB Library Lab.

This introductory post by Mahendra Mahey as part of Digital cultural heritage as a social resource project discusses how cultural institutions (GLAMs) need to go beyond merely digitising and providing access to the heritage they curate. Mahendra argues that GLAMs need to actively support computational access for the digital cultural heritage they curate and they need to encourage remixing, reuse, and experimentation to spark creativity with public engagement. 

He shows how initiatives around the world such as British Library Labs (2012-2021), the National Library of Scotland's Data Foundry,  Library of Congress Labs, the Metropolitan Museum of Art, the National Palace Museum of Taiwan and experiments in Estonia’s GLAMs show a “remixing mindset” around digital heritage. 

He argues that most of all staff within GLAMs need to adopt and encourage this remixing mindset to get their users to unlock the potential of the digital collections and data they curate.

Crossroads of Curiosity was an art installation of four lightpanel murals created by artist David Normal for Burning Man (2014) and the British Library (2015) that used Victorian Era book illustrations exclusively from the digitised collection of the British Library to create the artwork,

The DeCoGLAM (Decolonising GLAMs) project is building a community to decolonise GLAMs and other organisations and is led by Dr Milena Dobreva from the University of Strathclyde. 

DeCoGLAM acknowledges the colonial legacies of many GLAMs in the Global North, especially where eurocentric narratives have historically dominated collections and practices.  For the Global South, it means trying to reclaim agency over the cultural heritage that was displaced or misrepresented, while asserting local approaches to curating knowledge and cultural memory. For migrants and diasporic communities, it is about making GLAMs spaces for co-creation that can attempt to provide a rebalancing of what ownership, authorship, and representation means in today’s increasingly polarised world.

You can join the community via Linkedin and find out more information from the project website.

This working group was set up within the Europeana Research Community and EuropeanaTech Community and works to adapt and develop the concept of datasheets for the cultural heritage sector particularly for computational use.

The first version of datasheets for digital cultural heritage was described in this article and is available on Zenodo.

The second version for these datasheets were discussed at this webinar. The second version is also available on Zenodo.

Imaginary Cities was an arts-research project and exhibition by Michael Takeo Magruder that transformed the British Library’s online collection of historic urban maps into fictional cityscapes for the Information Age. 

The work explores how large repositories of digitised cultural materials can give rise to unique born-digital artefacts, real-time experiences and physical creations that are inspiring and relevant to contemporary audiences.

Dan Van Strien is Machine Learning Librarian at Hugging Face which is building an AI community collaborating on models, datasets and applications.  He regularly posts on Hugging Face as well as Linkedin.

Dan is trying to help democratise machine learning for the GLAM (Galleries, Libraries, Archives and Museums) sector and draw on ideas from library science to help develop a healthier machine learning ecosystem. 

Mario Klingemann is a German artist considered an early pioneer in the use of computer learning in the arts.  His works examine creativity, culture, and perception through machine learning and artificial intelligence, and has appeared at the Ars Electronica Festival, the Museum of Modern Art New York, the Metropolitan Museum of Art New York, the Photographers’ Gallery London, the Centre Pompidou Paris, and the British Library.

Using semi-automated image classification and machine learning techniques he was able add meaningful tags to the images and start the creation of thematic collections. Over the course of a year he tagged tens of thousands of images in the British Library Flickr collection and also added lots of collections to the "Lost Visions" project by the Cardiff University that is also using the BL material. Furthermore, he created various artworks with the found images - many of which are based on the principle of order. 

Created by researcher and artist David Zvi Kalman of Print-O-Craft Press and his collaborators, the Seder Oneg Shabbos bentsher, reimagines a traditional Jewish prayer book by blending historical artistry with contemporary inclusivity, incorporating over 120 images from sources like the British Library and texts for same-sex weddings and Yiddish songs. 

Originally designed for two weddings, it has since been used at hundreds of events, sold over 50,000 copies, and helped launch the Print-O-Craft publishing house. The project demonstrates how digital collections, especially those of the British Library, can make possible a high-quality, historically rich book that reconnects modern readers with the elegance of premodern Jewish print culture.

Jonathan Reeve works in the field of computational literary analysis, writing computer programs that help to understand novels and poetry.

For example, his project Git-Lit, inspired by GITenberg aims to convert British Library digital texts (in ALTO XML) into version-controlled, human-readable formats (like AsciiDoc), and host them on GitHub to allow collaborative correction, enhancement, and distribution (read more here). 

Learn more about other projects Jonathan has carried out from his personal website below.

Nick Cave is a designer, web developer, teacher and animator. He has been using digital cultural heritage to make a number of different animations especially celebrating history, through his Six Second History project. 

Six Second History started as a creative challenge to try and strip content from flat two dimensional Flickr British Library images and animate them in a three dimensional landscape to create a story which could then be viewed on social media. 

The Illustrated Newspaper Analytics project by Paul Fyfe and Qian Ge develops computer vision methods to study thousands of illustrations in 19th-century British newspapers, addressing a major gap in digital humanities research on visual materials. 

By analyzing pixel-level data rather than semantic content, the team identified patterns such as frequent maps, night scenes, facial portraits, and early halftone photographs, while also refining workflows to handle the technical challenges of historical images. Using British Library digitised newspapers, the project demonstrated how cultural analytics can reveal large-scale patterns in Victorian visual culture, while advancing interdisciplinary methods across literature, history, and engineering.

From 2016-2018 Always Already Computational: Collections as Data documented, iterated on, and shared current and potential approaches to developing cultural heritage collections that support computationally-driven research and teaching. 

With funding from the Institute of Museum and Library Services, Always Already Computational held two national forums, organized multiple workshops, shared project outcomes in disciplinary and professional conferences, and generated nearly a dozen deliverables meant to guide institutions as they consider development of collections as data. Collections as Data Interest Group is available at the research data alliance.

“Mouthful of Dust” is a cinematic web experience developed by State Library Victoria Lab, centered on 3D scans of artefacts related to Australian bushranger Ned Kelly. 

It lets users explore the objects in forensic detail—e.g. entering his suit of armour or inspecting engravings on his rifle—while accompanying the visuals with spatial audio and narrative contributions from Australian authors. 

The project demonstrates how digital technologies can not only preserve physical cultural heritage but also broaden access, enabling deeper engagement, remote exploration, and inclusive experiences.

The project funded by the Ministry of Education and Culture between 2024 and 2026 explores how Estonia’s digitized museum collections can serve as a social resource, helping communities adapt to change and fostering inclusive access to heritage.

It emphasizes shifting focus from mere preservation to active societal use, co-creation, and experimentation, by engaging diverse user groups and developing new applications of cultural heritage.

The project aims to produce at least three pilot digital heritage applications and a methodological “toolbox” to support participatory, people-centered heritage initiatives.

Digital infrastructure for Estonian humanities research or (HUMAL) aims to build a shared digital lab that consolidates scattered humanities data across archives, museums, and research institutions in Estonia to make it easier to discover, link, and use. 

It will also provide training materials and services (e.g. data analysis, visualization) to support researchers, educators, and cultural creators in working with digital humanities and cultural data. 

Initiated in July 2025 and led by the Estonian Literary Museum with partners like the Estonian National Museum, the initiative seeks to enhance collaboration and infrastructure in the Estonian humanities sector. 

The Institutional Data Initiative is a research initiative at Harvard Law School Library.  They work with knowledge institutions—from libraries and museums to cultural groups and government agencies—to refine and publish their collections as data to facilitate responsible AI training.

The Open GLAM Survey examines how GLAMs (Galleries, Libraries, Archives and Museums) make open access data – whether digital objects, metadata or text – available for re-use. 

Its working definition of ‘open’ is guided by Open Knowledge Foundation’s Open Definition. Its summary statement is ‘open means anyone can freely access, use, modify, and share for any purpose’ and the definition helpfully provides a list of licences, rights statements and legal tools that accord with this spirit.

Responsible AI in Libraries and Archives (2022-2025) was an IMLS-funded project aiming to support ethical decision-making for AI projects in libraries and archives. 

The project produced tools and strategies that support responsible use of AI in the field. AI projects in libraries and archives can support increased impact and new uses of resources, and Responsible AI provides new data, new resources, and new strategies that will prepare our profession for the methodical consideration of potential harms of AI projects. 

D-CRAFT stands for “Digital Content Reuse Assessment Framework Toolkit.” The toolkit helps measure use and reuse of digital objects. It can help streamline the reuse assessment process and make it more efficient.

The toolkit contains how to use 10 different assessment methods, an interactive tutorial for each method, how to tell stories of impact, how to write a takedown policy, Ethical guidelines, Related resources, Glossary of terms.

In 2013, the British Library released over a million images onto Flickr Commons for anyone to use, remix and repurpose.

"Moments" is a small collection of these images brought to life. 

At the time of creation Joe was a student, and received a 'First' for the project in his Motion Graphics module.

'Palimpsest: Telling Edinburgh’s Stories with Maps Through an Augmented Reality Phone Application' was created by Beatrice Alex, Miranda Anderson, Ian Fieldhouse, Claire Grover, David Harris-Birtill, Uta Hinrichs, James Loxley, Jon Oberlander, Nicola Osborne, Lisa Otty, Aaron Quigley, James Reid & Tara Thomson.

The project accessed over 380,000 full-text literary works from multiple providers, including the British Library’s entire Nineteenth Century Books Collection, from which over 5,500 relevant items were retrieved and 111 identified as Edinburgh-specific. British Library content represents more than 20% of the Palimpsest data, with excerpts linked through Historical Texts, and the team received strong support from BL Labs in data access and formatting.

The Digital Music Lab was an Arts Humanities Research Council (AHRC) funded initiative that aimed to bridge the methodological and infrastructural gap between musicology and Music Information Retrieval (MIR) by enabling large-scale, data-driven music research. 

It provided software infrastructure and prototype services to explore, analyse, and visualise large music collections (audio, scores, metadata), including collections held by the British Library and other major repositories. The project outputs included open data, open-source tools, and interactive interfaces to support quantitative and empirical research across diverse musical traditions. 

Janus Druz's entry for the British Library Labs Crowdsourcing Game Jam. The goal was to make a game that helps to crowdsource the collection of metadata about the digitised historical images of the British Library.

Tag Attack is a game developed for the British Library Crowdsourcing Game Jam. It explored the way games can be used to encourage participating in crowdsourcing projects, i.e. tagging items from a digital collection like the British Library Flickr collection. It was developed by Antonio Jesús Sánchez Padial.

Flickr is one of the largest cultural archives in human history with millions of new images, likes and comments added each day. The Flickr Foundation created the Data Lifeboat, a user-friendly archiving solution to ensure memories on Flickr can be enjoyed by future generations, in easily browsable packages, e.g. the Flickr Life Boat for the Enola Gay.

Kris Hofmann and Claudia Rosa Lukas created a 80 second animation in ultra high res 4k as well as 5 vines to accompany the Austrian contribution to the International Fashion Showcase London, which is organised by the British Council every year. The show was on at Somerset House for two weeks and themed 'Fashion Utopias'.

Stretching the boundaries of the creative use of public domain materials, graphic artist Paul Rand Pierce scoured the over one million images in the British Library's Flickr Photostream of public domain works, and integrated the hidden gems of the collection into a unique, illustrated work of fiction entitled 'Realmland: The Imaginary Kingdom of Maynard Tate'.

Directed by Ling Low, 'Hey There, Young Sailor' is a music video for the band 'Impatient Sisters' using images from the British Library.  It mixes live action, archive images, digital animation and hand drawn illustrations, the film tells an enchanting tale of two lovers lost at sea. 

British Library Art Project, animated 2D Illustrations from the British Library Flickr Commons Collection by Jiayi Chong. This project recreates a fictional scene using painted people and animals from the British Library archive and brings them to life using the Creature Animation Tool. Characters exhibit realistic movement, with cloth and hair that sways naturally with motion/wind.

Poetic Places was a free app for the iPhone Operating System (iOS) and Android devices, and the main outcome of a collaboration between Sarah Cole of TIME/IMAGE and the British Library.

Poetic Places let users uncover hidden histories and culture, bringing poetic depictions of places into the everyday world, helping them to encounter poems and literature in the locations described, accompanied by audio visual materials drawn from archive collections.

British Library Labs was established in 2012 to encourage people to experiment with the British Library's digital collections and data. Their YouTube channel contains 182 videos relating to the project.  

Filipe Bento has been publishing a curated daily AI Newsletter on LinkedIn and contains over 400 postings with over 1000 subscribers. It's a great way to stay on top of the news and trends of AI in the world and to learn about amazing new discoveries, new services and tools.

Curated by AI expert Alejandro Saucedo who advises the United Nations on AI has been curating the Machine Learning Engineer with nearly 200 editions and 350 issues and over 55,000 subscribers on LinkedIn for Machine Learning professionals and enthusiasts. They receive weekly articles & tutorials on production Machine Learning & Machine Learning Operations.

Curated and released by the Austrian National Library Labs the Austroflug dataset comprises of about 4,800 digitized aerial photographs of Austria taken between 1930 and 1935, accompanied by descriptive metadata. Through a crowdsourcing initiative, users helped enrich the dataset by adding geographical coordinates for many images, and a Jupyter Notebook documenting the georeferencing method is available.

The dataset is openly accessible and integrated with an interactive map and search interface to allow exploration by location or image signature.

The artspace is a platform to showcase artworks engaging with the data collections provided by ONB Labs (the Austrian National Library) .

The Carpentries is a global non-profit organisation that teaches foundational coding and data science skills to researchers through open, community-developed lessons and workshops.  Their mission is to empower a diverse, collaborative community to conduct efficient, reproducible research by providing inclusive, evidence-based training and open educational resources. 

It consists of 3 carpentries that build lesson programs teaching:
1 - Data Carpentry - universal data skills.
2- Library Carpentry - software and data skills for people working in library and information-related roles.
3 - Software Carpentry - basic lab skills for research computing

Programming Historian is a volunteer-driven, open access project that publishes peer-reviewed tutorials on digital humanities methods, using open source tools and making materials freely and widely available under CC-BY licensing. 

It follows a rigorous editorial workflow to ensure clarity and accessibility for non-specialist readers, and does not charge publication or access fees (diamond open access). 

Its governance emphasizes diversity and inclusion, and it has expanded over time to include multilingual publications in English, Spanish, French, and Portuguese.

DOORS, launched by Ars Electronica, MUSEUM BOOSTER, and Ecsite, is a digital incubator designed to support small and medium-sized museums. It helps them build sustainable digital strategies and projects by offering mentorship, tools, and knowledge exchange. The program aims to bridge the digital capacity gap between smaller and larger cultural institutions.

The Balanced Value Impact (BVI) Model is a framework for assessing the impact of digital resources, especially in cultural, heritage, academic, and creative sectors, by combining insights from multiple impact assessment disciplines into a cohesive process., developed by Professor Simon Tanner.

It defines impact as “the measurable outcomes arising from the existence of a digital resource that demonstrate a change in the life or life opportunities of the community” and guides organisations through five functional stages: context, design, implementation, outcomes and results, and review/respond. 

The model encourages balancing economic, social, cultural, and internal organizational perspectives to create a more nuanced evidence-based narrative of value and change.

The CORAL project investigates how to train and use large language models (LLMs) under strict legal, technical, and quality constraints, focusing on transparency, lawful data use, and retrieval-augmented generation (RAG). 

Drawing on exclusive data sources such as the German National Library, Internet Archive, Common Crawl, and finance sector partners, the project explores how obfuscation methods can make sensitive or restricted data usable for AI without breaching regulations. 

Its goal is to ensure professional LLM applications generate accurate, source-based, copyright-compliant texts, supported by new RAG methods that both enrich outputs and protect training data.

KorAP is a next-generation corpus analysis platform designed for managing very large, multiply annotated corpora and supporting complex, methodologically valid linguistic research. 

Developed as a successor to COSMAS I and II, it addresses new challenges in corpus linguistics such as handling rapidly growing datasets, enabling reproducibility and collaborative research, and integrating with distributed infrastructures like CLARIN. 

The platform enhances user experience with advanced features including metadata-driven virtual corpus creation, improved query syntax for multi-layer annotations, faster processing, and graphical result displays.

SANE is a secure analysis environment developed by SURF that enables researchers to work with sensitive data while data providers retain full control over access and output. 

It offers two modes: Tinker (interactive virtual desktop use without export of raw data) and Blind (non-interactive analysis via scripts or containers where the researcher never sees the raw data). 

The environment is certified (ISO-27001), undergoes penetration testing, and is available via the SURF Research Cloud to streamline secure collaboration on sensitive datasets.

A demonstration webinar is available.

Ollama is an open-source tool that simplifies running and managing large language models (LLMs) on a local machine, providing data privacy, offline access, and reduced latency by avoiding cloud reliance. 

It acts as a local model server with a command-line interface (CLI) and supports a wide range of open-source models, allowing users to download, manage, and integrate them into various projects without deep technical expertise. 

Open WebUI is a self-hosted, extensible interface for running large language models (LLMs) locally or offline, supporting both Ollama models and OpenAI-compatible APIs. 

It includes features such as granular permissions, user groups, Markdown/LaTeX support, responsive design, and built-in RAG (retrieval-augmented generation) capability. 

Open WebUI can be deployed via Docker, Kubernetes, or Python (e.g. via pip or uv), ensures local data storage and privacy, and aims for ease of use with offline capabilities. 

LiteLLM is an open-source Large Language Model (LLM) gateway and Python SDK that lets you call over 100 different large language model APIs (e.g. OpenAI, Anthropic, Hugging Face, Azure) via a unified OpenAI-compatible interface. 

It provides features like request fallback, cost & usage tracking, rate limiting, and standardized input/output formats.

The LC Labs AI Framework is a planning framework developed by the Library of Congress to guide responsible experimentation and adoption of AI technologies in cultural heritage institutions. 

It is structured around three phases — Understand, Experiment, and Implement — and considers three key dimensions: data, models, and people, using worksheets, risk assessments, and stakeholder engagement to inform decisions. 

The framework is shared openly (CC0 1.0) to encourage adoption, feedback, and adaptation by other organizations. 

Liberating Structures is a collection of 33 simple, inclusive methods designed to transform how people interact in groups by redistributing control and enabling everyone’s contribution. 

The approach emphasizes minimal structure that unlocks creativity, engagement, and innovation in meetings, planning, and decision-making. 

 It is intended to replace or complement conventional practices like presentations or brainstorming by embedding more participatory, dynamic processes. 

Wikidata is a free and collaborative knowledge base that stores structured, linked data to support Wikipedia, Wikimedia projects, and external applications. It functions as a central repository of facts—such as people, places, events, and concepts—that can be accessed and reused across languages and platforms.

Maintained by a global community, Wikidata enables consistent, open, and machine-readable data for research, cultural heritage, and digital innovation.

Wikimedia Commons is a free, online repository of images, videos, sounds, and other media files that anyone can use, share, and contribute to. It serves as the central media archive for all Wikimedia projects, including Wikipedia, making content easily accessible across languages and platforms. 

All files are freely licensed or in the public domain, encouraging open knowledge sharing and creative reuse worldwide.

The Mechanical Curator was an experimental project by the British Library Labs, created by Ben O'Steen, that provided random, interesting images from the Library's 17th to 19th-century digitized book collections via a Tumblr account and later uploaded approximately one million images to Flickr Commons. 

Its purpose was to offer a unique, "undirected" experience, allowing users to discover and explore the Library's vast public domain materials.

Microsoft’s AI for Good Lab is a research group within Microsoft Research that develops and applies AI solutions to tackle global challenges in health, sustainability, equity, and more. 

It works via initiatives such as LINGUA (for underrepresented European languages) and projects like SPARROW (for biodiversity monitoring) and climate-risk mapping. 

The Lab emphasizes open data, collaboration with institutions and communities, and ethical use of AI for social and environmental impact. The ideas behind this can be read in the 'AI For Good' book.

Funded by the Getty Conservation Institute, Arches is an open-source data management platform that is freely available for organizations worldwide to install, configure, and extend in accordance with their individual needs and without restrictions on its use. 

Due to the complex and varied nature of cultural heritage data, and to promote interoperability and sustainable data practices,.

Arches has been developed as a standards-based, comprehensive and flexible platform that supports a wide array of uses, including Arches for Science, and involves a diverse international community of users.

The Banned Books Museum in Tallinn, Estonia run by Joseph Dunnigan it collects and exhibits books from around the world that have been banned, censored, or burned, with a permanent collection of over 400 books from more than 100 countries. 

It aims to educate visitors about the many forms of literary censorship while remaining politically neutral and encouraging visitors to form their own understanding of censorship.